PackLab: A Comprehensive Framework for Developing, Training, and Evaluating MLLMs in Robotic Bin Packing
PackLab releases a specialized multimodal model and benchmark for closed-loop robotic bin packing.
PackLab is a framework for developing, training, and evaluating multimodal large language models on closed-loop robotic bin packing. PackLab-Suite is a physics simulator for generating packing trajectories, PackLab-VLM selects objects and predicts placements from evolving container state, and PackLab-Bench offers standardized scenarios at several difficulties. Experiments report that PackLab-VLM outperforms hand-crafted geometric heuristics, traditional reinforcement learning, and general-purpose MLLMs across object sets and container configurations. Code, model, dataset, and benchmark are released on GitHub.
- PackLab targets closed-loop sequential decisions for robotic bin packing.
- PackLab-Suite simulates diverse trajectories and physical packing outcomes.
- PackLab-VLM jointly selects objects and predicts placements from evolving state.
- PackLab-Bench standardizes scenarios at multiple difficulty levels.
- The specialized MLLM beats heuristics, traditional RL, and general MLLMs on average.
Full article188 words · extracted from huggingface.co · click to collapse
Robotic bin packing requires long-horizon sequential decision-making, as each object placement affects the available space for subsequent packing. Existing methods primarily rely on hand-crafted geometric heuristics that optimize predefined objectives or reinforcement learning policies learned through trial and error over predefined training configurations. Despite recent advances in multimodal large language models (MLLMs) for this task, their potential for closed-loop sequential decisions across heterogeneous packing configurations remains underexplored. To address this gap, we introduce PackLab, a comprehensive framework for developing, training, and evaluating MLLMs for closed-loop robotic bin packing. PackLab-Suite provides a physics-based simulation platform for scalable generation of diverse training packing trajectories and evaluation of their physical outcomes. PackLab-VLM is a packing-specialized MLLM that understands the evolving object and container states to jointly select objects and predict placements in a closed-loop manner. PackLab-Bench provides standardized packing scenarios at multiple difficulty levels for systematic evaluation. Extensive experiments demonstrate that, on average, PackLab-VLM outperforms conventional packing heuristics, traditional reinforcement learning methods, and general-purpose MLLMs across object sets and container configurations, highlighting the potential of MLLMs for long-horizon robotic packing. The code, model, dataset, and benchmark are available at https://github.com/Correr-Zhou/PackLab .
Text extracted automatically; images, tables and formatting may be missing. Original: https://huggingface.co/papers/2609.23784